NDT Advance Access originally published online on October 13, 2006
Nephrology Dialysis Transplantation 2007 22(1):9-11; doi:10.1093/ndt/gfl580
© The Author [2006]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
Which cardiovascular risk factors matter in chronic kidney disease?
Colin Baigent and
Martin Landray
Clinical Trial Service Unit and Epidemiological Studies Unit (CTSU), Oxford, UK
Correspondence and offprint requests to: Prof Colin Baigent, Clinical Trial Service Unit and Epidemiological Studies Unit (CTSU), Richard Doll Building, Old Road Campus, Roosevelt Drive, Oxford OX3 7LF, UK. Email: colin.baigent{at}ctsu.ox.ac.uk
Keywords: blood pressure; cholesterol; epidemiology; haematocrit; randomized trials; risk factors
 |
Why measure risk factors for vascular disease?
|
|---|
Identifying cardiovascular risk factors among patients with
chronic kidney disease (CKD) is potentially useful for two main
reasons. First, such risk factors may be used to predict the
future risk of an event, which may help physicians to plan treatment.
Second, they may be targets for intervention to prevent such
events. It is now clear that a large number of risk factors
independently predict the risk of cardiovascular outcomes among
dialysis patients [
1] and among patients with lesser degrees
of renal impairment [
2]. But, although a risk score based on
independent risk factors can be assembled without particular
regard to whether they are causal, treatments which modify a
particular risk factor will only be effective for the prevention
of vascular disease if that risk factor is a cause of such disease.
|
Limitations of dialysis studies for identifying causes of vascular disease
|
|---|
The starting point for identifying risk factors that might prove
to be possible targets for intervention in CKD is population-based
observational studies. But we face an immediate difficulty when
evaluating observational studies among dialysis patients because
such studies are especially prone to confounding, thereby distorting
epidemiological associations. The problem is that dialysis patients
have a complex array of different metabolic disorders, comorbid
conditions and treatments, so that characterizing them with
a limited number of measurements is virtually impossible. Even
after correction for measured confounding variables, therefore,
substantial ‘residual’ confounding is likely to
be present. An illustration of this difficulty is provided by
the phenomenon of ‘confounding by disease’, or ‘reverse
causality’. Disorders such as malnutrition, for example,
may influence the metabolism of risk factors so profoundly that
genuinely causal associations are obscured. If we take the example
of cholesterol [or more specifically low-density lipoprotein
(LDL) cholesterol], which is known to be an important cause
of atherosclerosis, negative associations have been observed
between cholesterol and mortality in epidemiological studies
among people who are very elderly [
3], or who have advanced
heart failure [
4], cancer [
5] or AIDS [
6]. We find exactly the
same phenomenon in dialysis patients [
7,
8]. These negative associations
should not, however, be taken as evidence that higher cholesterol
protects patients from atherosclerotic disease in these conditions.
Instead, a more plausible explanation is that down-regulation
of hepatic lipoprotein production in those at highest risk of
death results in low cholesterol levels [
9]. Evidence that such
patterns are most likely the result of confounding comes from
studies in the very elderly, in which the negative associations
in observational studies [
3] are strongly refuted by the unconfounded
evidence from large-scale randomized trials. A meta-analysis
of 90 000 patients in 14 statin trials has shown clearly that
each 1 mmol/l reduction in LDL cholesterol resulted in a one-fifth
reduction in major vascular events [myocardial infarction or
death from coronary heart disease (CHD), stroke or coronary
revascularization] in those aged 75 or over (
P = 0.0001) [
10].
|
Alternative strategies for identifying causes of vascular disease in CKD
|
|---|
A more reliable strategy for identifying risk factors relevant
to vascular disease in patients with CKD may be to extrapolate
information from observational studies conducted in non-renal
populations. Several aetiologically distinct vascular pathologies
are found among patients with CKD, including atherosclerosis,
vascular stiffness, calcification and left-ventricular hypertrophy.
Registry studies have shown, for example, that left-ventricular
hypertrophy is present in about three-quarters of patients commencing
dialysis [
11] and congestive heart failure is the most common
recorded type of morbidity for inpatients with CKD [
12]. Atherosclerotic
conditions such as myocardial infarction and stroke account
for about one-third of vascular deaths [
12]. Of these pathologies,
left-ventricular hypertrophy and atherosclerosis occur commonly
in non-renal populations, and observational studies have identified
potentially causal risk factors for them. Whilst the
strength of any such association might differ in CKD, it is unlikely
to be completely abolished. Hence, in renal populations, randomized
trials to test treatments that modify risk factors identified
in non-renal populations should be a starting point in the effort
to prevent vascular disease in CKD.
|
Average levels of risk factors may to be too high for CKD patients
|
|---|
In non-renal populations, blood LDL cholesterol is positively
and continuously associated with the risk of CHD across a wide
range of cholesterol levels, with no increase in risk of CHD
at low LDL concentrations. Randomized trials in non-renal populations
have shown that lowering LDL cholesterol reduces this risk [
10],
thus demonstrating that LDL cholesterol is a cause of CHD. Since
the absolute risk of CHD is higher among CKD patients than among
those without CKD, if we were able to correct completely for
confounding then we might hypothesize that the association between
usual LDL cholesterol and CHD would run parallel but higher
than the association in a non-renal population (
Figure 1). This
implies that a large absolute reduction in CHD may be possible
by reducing LDL cholesterol among patients with CKD even in
those with ‘normal’ (i.e. average) LDL cholesterol
levels. For LDL-cholesterol, the key point is that the level
of the risk factor does not have to be ‘abnormal’
to be a potentially important target for intervention: all that
matters is that the risk factor is causal, and that the causal
association is continuous right down to the lowest levels that
can be achieved safely. (This is not intuitive, because as physicians
we are trained to correct abnormal levels, and leave normal
levels well alone.) Since there are no known hazards of reducing
LDL cholesterol to very low levels, at least in non-renal populations
[
10], the ongoing randomized trials of cholesterol-lowering
treatments in patients with CKD (SHARP [
13] and AURORA [
14])
do not have any lower threshold for LDL cholesterol in their
eligibility criteria.
A similar argument can be applied to blood pressure, since observational
studies in non-renal populations have demonstrated positive
associations between blood pressure and the risk of CHD, stroke
and congestive heart failure [
15], and randomized trials of
blood pressure–lowering regimens have shown that lowering
blood pressure reduces the risk of these outcomes [
16,
17]. Hence,
it may be hypothesized that lowering blood pressure might be
beneficial to patients with CKD even when blood pressure control
is regarded as adequate, provided such reductions can be tolerated
by patients. Testing this hypothesis among patients with CKD
is especially important because raised blood pressure is associated
with three important vascular outcomes (CHD, stroke and heart
failure) as well as progression of renal disease, so the potential
absolute benefits of further reductions in blood pressure in
high-risk patients with CKD who have ‘well controlled’
blood pressure are substantial.
|
Assessing risk factors which are specific to CKD
|
|---|
Some risk factors which have been implicated in the pathogenesis
of vascular disease in CKD may be difficult to study in observational
studies among non-renal patients because the range of usual
values in non-renal patients does not extend to the values which
occur in advanced CKD. For example, the usual values for haematocrit
in the general population are
39–49% in men and 33–43%
in women [
18], but haematocrit falls below these ranges once
CKD becomes established. Observational studies examining the
association between haematocrit and CHD have been performed
in healthy people, and a meta-analysis of 16 such studies showed
that, as compared with those in the bottom tertile of haematocrit
values (mean 41.7%), the relative risk for CHD among those in
the top tertile (mean 46.3%) was 1.16 (95% CI 1.05–1.29)
[
19]. But, in contrast to studies of cholesterol or blood pressure,
these studies do not tell us very much about the nature of any
association in people with much lower haematocrit levels, such
as those prevailing in patients with CKD, in whom it is likely
that low haematocrit is associated with higher risk. In these
circumstances, the lack of relevant information from non-renal
observational studies and the potential for residual confounding
in observational studies of dialysis patients [
20] makes randomized
trials of erythropoietin essential if we are to determine whether
low haematocrit is a cause of vascular disease in CKD.
|
Summary
|
|---|
Risk factors for vascular disease in CKD can be useful both
for assessing the level of risk of vascular disease and as potential
targets for the prevention of such disease. Only those risk
factors which cause a major component of vasculopathy in CKD
are likely to be worth modifying. Owing to unavoidable residual
confounding, observational studies in dialysis patients (or
patients with advanced CKD) may provide misleading information
about the strength and direction of associations. Better control
of this confounding may be possible for particular risk factors
in non-renal populations, but ultimately randomized trials will
be essential, both for establishing the causal relevance of
candidate risk factors for vascular disease in renal failure,
and for determining the size of any clinical benefit.
Conflict of interest statement. None declared.
|
References
|
|---|
- Sarnak MJ and Levey AS. (2000) Cardiovascular disease and chronic renal disease: a new paradigm. Am J Kidney Dis 35:S117–S131.[ISI][Medline]
- Jungers P, Massy ZA, Khoa TN, et al. (1997) Incidence and risk factors of atherosclerotic cardiovascular accidents in predialysis chronic renal failure patients: a prospective study. Nephrol Dial Transplant 12:2597–2602.[Abstract/Free Full Text]
- Corti M-C, Garalnik JM, Salive ME, et al. (1997) Clarifying the direct relation between total cholesterol levels and death from coronary heart disease in older persons. Ann Intern Med 126:753–760.[Abstract/Free Full Text]
- Horwich TB, Hamilton MA, Maclellan WR, Fonarow GC. (2002) Low serum total cholesterol is associated with marked increase in mortality in advanced heart failure. J Cardiac Fail 8:216–224.[CrossRef][ISI][Medline]
- Chao FC, Efron B, Wolf P. (1975) The possible prognostic usefulness of assessing serum proteins and cholesterol in malignancy. Cancer 35:1223–1229.[CrossRef][ISI][Medline]
- Neaton JD and Wentworth DN. (1997) Low serum cholesterol and risk of death from AIDS. AIDS 11:929–930.[ISI][Medline]
- Degoulet P, Legrain M, Reach I, et al. (1982) Mortality risk factors in patients treated by chronic haemodialysis. Nephron 31:103–110.[ISI][Medline]
- Lowrie EG and Lew NL. (1990) Death risk in hemodialysis patients: the predictive value of commonly measured variables and an evaluation of death rate differences between facilities. Am J Kidney Dis 15:458–482.[ISI][Medline]
- Baigent C, Burbury K, Wheeler D. (2000) Premature cardiovascular disease in chronic renal failure. Lancet 356:147–152.[CrossRef][ISI][Medline]
- Cholesterol Treatment Trialists’ (CTT) Collaboration. (2005) Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet 366:1267–1278.[CrossRef][ISI][Medline]
- Foley RN, Parfrey PS, Harnett JD, et al. (1995) Clinical and echocardiographic disease in patients starting end-stage renal disease therapy. Kidney Int 47:186–192.[ISI][Medline]
- US Renal Data System. (2005) Annual Data Report: atlas of end-stage renal disease in the United States. National Institutes of Health. (National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD).
- Baigent C and Landray MJ. (2003) Study of Heart and Renal Protection (SHARP). Kidney Int 63:Suppl 84, S207–S210.[CrossRef]
- Fellstrom B, Zannad F, Schmieder R, et al. (2005) Effect of rosuvastatin on outcomes in chronic haemodialysis patients—design and rationale of the AURORA study. Curr Control Trials Cardiovasc Med 6:9.
- Prospective Studies Collaboration. (2002) Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet 360:1903–1913.[CrossRef][ISI][Medline]
- Law M, Wald N, Morris J. (2003) Lowering blood pressure to prevent myocardial infarction and stroke: a new preventive strategy. Health Technol Assess 7:1–94.[Medline]
- Psaty BM, Smith NL, Siscovick DS, et al. (1997) Health outcomes associated with antihypertensive therapies used as first-line agents. A systematic review and meta-analysis. JAMA 277:739–745.[Abstract]
- Laposata M. (1992) The New England Journal of Medicine SI Unit Conversion Guide , Boston, MA.
- Danesh J, Collins R, Peto R, Lowe GDO. (2000) Haematocrit, viscosity, erythrocyte sedimentation rate: meta-analyses of prospective studies of coronary heart disease. Eur Heart J 21:515–520.[Free Full Text]
- Foley RN, Parfrey PS, Harnett JD, et al. (1996) The impact of anemia on cardiomyopathy, morbidity, and mortality in end-stage renal disease. Am J Kidney Dis 28:53–61.[ISI][Medline]
Received for publication: 25. 7.06
Accepted in revised form: 24. 8.06
CiteULike 
Connotea 
Del.icio.us What's this?
Top
Why measure risk factors...
Limitations of dialysis studies...